UTF-8 problem in python when reading chars

Question

I'm using Python 2.5. What is going on here? What have I misunderstood? How can I fix it?

in.txt:

Stäckövérfløw

code.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-
print """Content-Type: text/plain; charset="UTF-8"\n"""
f = open('in.txt','r')
for line in f:
    print line
    for i in line:
     print i,
f.close()

output:

Stäckövérfløw

S t � � c k � � v � � r f l � � w

Answer 1

Check this out:

# -*- coding: utf-8 -*-
import pprint
f = open('unicode.txt','r')
for line in f:
    print line
    pprint.pprint(line)
    for i in line:
        print i,
f.close()

It returns this:

Stäckövérfløw
'St\xc3\xa4ck\xc3\xb6v\xc3\xa9rfl\xc3\xb8w'
S t ? ? c k ? ? v ? ? r f l ? ? w

The thing is that the file is just being read as a string of bytes. Iterating over them splits the multibyte characters into nonsensical byte values.

Answer 2

Use codecs.open instead, it works for me.

#!/usr/bin/env python
# -*- coding: utf-8 -*-
print """Content-Type: text/plain; charset="UTF-8"\n"""
f = codecs.open('in','r','utf8')
for line in f:
    print line
    for i in line:
        print i,
f.close()

Answer 3

for i in line:
    print i,

When you read the file, the string you read in is a string of bytes. The for loop iterates over a single byte at a time. This causes problems with a UTF-8 encoded string, where non-ASCII characters are represented by multiple bytes. If you want to work with Unicode objects, where the characters are the basic pieces, you should use

import codecs
f = codecs.open('in', 'r', 'utf8')

If sys.stdout doesn't already have the appropriate encoding set, you may have to wrap it:

sys.stdout = codecs.getwriter('utf8')(sys.stdout)

Answer 4

print c,

Adds a "blank charrecter" and breaks correct utf-8 sequences into incorrect one. So this would not work unless you write a signle byte to output

sys.stdout.write(i)